Users Guide to the QDE Toolkit Pro Excel Tools for Presenting Metrological Comparisons by B'M' Wood, - PowerPoint PPT Presentation

1 / 15
About This Presentation
Title:

Users Guide to the QDE Toolkit Pro Excel Tools for Presenting Metrological Comparisons by B'M' Wood,

Description:

User's Guide to the QDE Toolkit Pro' Excel Tools for ... 14. VBA Source Code 183 - 192. 15. Afterthoughts 193 - 202. National Research Conseil national ... – PowerPoint PPT presentation

Number of Views:187
Avg rating:3.0/5.0
Slides: 16
Provided by: robdouglas
Category:

less

Transcript and Presenter's Notes

Title: Users Guide to the QDE Toolkit Pro Excel Tools for Presenting Metrological Comparisons by B'M' Wood,


1
Users Guide to the QDE Toolkit Pro Excel
Tools for Presenting Metrological Comparisons
byB.M. Wood, R.J. Douglas A.G. Steele Guide
Version 2.07 for QDE Toolkit Pro Version 2.07
May 2002September 2003
National Research Conseil national Council
Canada de recherches
2
2
Table of Contents for this Guide
The detail and screenshot resolution make this a
large file. Sometimes it needs to be broken into
Chapter files. Chapter Pages 1.
Introduction/Overview 3 - 15 2. Getting
Started 16 29 3. Building Equivalence
Tables (i) 30 - 43 4. Equivalence Tables
(ii) 44 - 58 5. Building Confidence-in-CMC
Tables 59 - 82 6. Degrees of Freedom 83
- 93 7. Covariance and Correlation
Coefficients 94 - 102 8. Graphing and Pooling
(i) 103 - 110 9. Graphing and Pooling (ii)
111 - 128 10. Project Organization
Excel/Workbooks/Worksheets 129 - 143 11.
Admitting and Excluding Labs in the Pool -
Outliers 144 - 148 12. Automated RVs and RV
creators 149 - 176 13. Creating Lab Views
177 - 182 14. VBA Source Code 183 -
192 15. Afterthoughts 193 - 202
3

Ch 1 3
May 2, 2002
National Research Conseil national Council
Canada de recherches
Users Guide to the QDE Toolkit Pro
Excel Tools for Presenting Metrological
Comparisons byB.M. Wood, R.J. Douglas A.G.
Steele
Chapter 1. Introduction and Overview
In this chapter, we touch on the role of the
Toolkit. An inter-Laboratory comparison, after
perhaps difficult analysis into differences and
uncertainties, looks deceptively simple. Choose a
reference value, assign its uncertainty, apply
standard null hypothesis tests and youre
done It is often much more difficult than this,
and at times is even acrimonious.The QDE
Toolkit Pro includes Excel macros to deal with
Equivalence, with or without the intermediary of
a particular reference value (RV), and provides
facilities for examining various candidate RVs
and u(RV)s.
4
Ch 1 4
Whats this QDE?
It is now only a minor part of the tools
available in the current version of the QDE
Toolkit. Its a single number, having the units
of the measurement, that can report how well two
measurement processes agree, based on a
comparison that has revealed a difference and an
uncertainty. It is the Quantified Demonstrated
Equivalence of the comparison, and is
appropriate only where a single parameter is
needed to summarize differences to be expected
from similar measurements.
5
Ch 1 5
Whats the use of a QDE number?
QDE supports simple sentences... On the basis
of this comparison, similar measurements can be
expected to agree within 2.3??m, 19 times out of
20. or ... , with 95 confidence. QDE is
able to blend the results from the two
measurement processes into one QDE parameter. It
converts four to seven parameters into one
easy-to-use parameter, usually starting from two
values with two uncertainties, and perhaps two
effective degrees of freedom and a correlation
coefficient.
6
Ch 1 6
Whats the QDE Toolkit?
Its a suite of Excel macros that can
automatically convert a small table of results,
from a comparison of N Labs measurement
processes, into a standard form such as the
Degrees of Equivalence Table with 2N2 entries.
QDE tabulations are also available in the
Toolkit. Version 1 of the QDE Toolkit was
written by A.G. Steele of NRCs Thermometry
Group, it was released for public use in October,
2000 and about 100 copies of Version 1.14 (and
its maintenance release 1.15) have been
distributed. The QDE Toolkit Pro is the next
major public release.
7
Ch 1 7
Whats New in the Toolkit Pro?
Loaded by Microsoft Excel, QDE Toolkit Pro
simply handles Key Comparisons of many Labs with
degrees of freedom and inter-laboratory
correlation coefficients, and provides simple
defaults if they are not available.
Now 10000 lines of Visual Basic code
and comments provide help with the
repetitive tasks of analysis and presentation
titles, units, comments and graphs that go beyond
error bars.
The toolkit displays all its work openly on the
Excel spreadsheets, rather than having macros
communicate with each other without your
supervision. It can also be shared by other open
workbooks.
8
Ch 1 8
Who should consider the Toolkit Pro?
It is for pros in metrology, who are familiar
with degrees of freedom and interlaboratory
correlation coefficients, and want to treat them
with rigor.
It is for those with a working knowledge of Excel.
It is for those who want to think about their
comparisons...
and communicate their thoughts clearly and
simply to others.
It still does all the simple analyses of Version
1.14, but explicitly writes out its default
assumptions infinite degrees of freedom and no
correlations...
9
Ch 1 9
Who should try the Toolkit Pro?
q Pilots for Key Comparisons the Toolkit
provides KCRV-independent analyses, can help
examine a variety of candidate KCRVs, and
automatically builds standard Equivalence Tables
for cross-checking with those done manually or
with other software. q Participants in Key
Comparisons who wish to present a rigorous
argument about some aspect of the Comparison. q
Those voting on MRA Appendix C CMCs, or those
providing counsel, when there is a similar Key
Comparison in Appendix B. q Those who want to
interpret, think about and justify results from
multi-laboratory comparisons. Here, the toolkit
computes, records, and prepares presentation
tables and graphs - and the expert is free to
have fun! The Toolkit Pro can also deal with
proficiency demonstrations effectively, but its
real strength is in dealing with comparisons
where the right answer is not easily apparent.
10
Ch 1 10
Why such a Toolkit?
As metrologists, how are we to cope with
reviewing and discussing the many, many
comparisons that are planned in support of the
CIPM MRA?
We will be examining candidate KCRVs, tabulating
MRA degrees of equivalence, and documenting
decisions made for the MRA Appendix C Repetitive
parts of all these tasks should be automated.
This toolkit is the authors response to these
challenges. It is offered in the belief that
others may find it useful.
11
Ch 1 11
About the Present Toolkit
The Toolkit runs within Microsoft Excel to
automate some of the rigorous analysis and
presentation required for the international
comparisons that are proliferating in metrology.
It calculates confidence, and confidence
intervals, in the QDE (Quantified Demonstrated
Equivalence) formalism. It works with Lab values
and reference values (RVs) that may have
unavoidable correlations. It deals with pooled
distributions and outliers. (Wood and Douglas,
Metrologia 35, 187-196 (1998)) (Steele, Wood and
Douglas, Metrologia 38, 483-488 (2002)) (Hill,
Steele and Douglas, Metrologia 38, accepted to
appear (2002)) Even those who choose not to use
the QDEformalism will like some other features
of the Toolkit. Presentation of comparisons in
forms suitable for MRA submissions and decisions
are included. Some functions are included that
are not documented in the Users Guide
2003-09-05, but rely on Excels documentation
tools.
12
Ch 1 12
What we see the Toolkit becoming
New releases of the Toolkit Pro have increased
the rigor and variety of functions and macros.
They more fully automate the work of calculating,
interpreting and presenting comparison data in
the useful formats that we have encountered.
As further techniques of analysis and
presentation aredeveloped and/or become popular,
we hope to automatethose as well. We encourage
you to suggest or contribute extensionsto this
toolkit to make it better suit your needs.
13
Ch 1 13
What are the Toolkits limitations?
Limitations arise from trying to be universal and
generic q To be universal, and target all fields
of measurement, it will remain a post-processor
to be used after most of the analysis specific to
a measurement field has been done. q To be
generic (the measurand was) rather than
specific ( on the measurement is predicted to
be), the Toolkit Pro assumes that time
dependences of the measurand have been removed
and that it does not have to deal with
time-dependent uncertainties.
14
Ch 1 14
Toolkit Distribution
Rather than distributing the Toolkit and Guide
yourself, please ask other potential users to
register by sending an email to
QDE_at_nrc-cnrc.gc.ca By registering with us, we can
notify you of any errors, flaws or problems which
have been identified. Our website
http//inms-ienm.nrc-cnrc.gc.ca/QDE maintains the
revision history, known problems, FAQs, other
documentation, papers and presentations. We plan
to coordinate Toolkit extensions to address other
specific common needs and provide a forum for
validation discussions.
15
Ch 1 15
About the Toolkit Users Guide
We are still in the process of improving and
enhancing both the QDE Toolkit and this Users
Guide. The QDE Toolkit Pro has been in use
outside NRC since early 2002. In order to avoid
problems associated with having different
versions in use, we ask you not to distribute
this Toolkit or the Guide to others. We will
provide the latest version of this software and
documentation to anyone who registers. This
Users Guide is available as a Microsoft
PowerPoint (.ppt) file or as an Adobe Acrobat
(.pdf) file. Because of the size of these files,
individual Chapter files are also available for
convenience in e-mailing and downloading.
Write a Comment
User Comments (0)
About PowerShow.com